• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

NVIDIA GTX 1080 Ti Confirmed; 980 Ti Owners to Enjoy "Step-up" Program

Exactly how is Pascal boring?
Pascal's desktop GPUs are more or less Maxwell on a smaller manufacturing process node. And Maxwell is also not that much different from Kepler. Volta will be the first major architectural redesign since Kepler. That's why.
 
Pascal's desktop GPUs are more or less Maxwell on a smaller manufacturing process node. And Maxwell is also not that much different from Kepler. Volta will be the first major architectural redesign since Kepler. That's why.

There's also from a purely enthusiast view, that Pascal offers nothing in DX12 and Vulkan. New tech (even API's) get us excited, and Pascal not doing anything there makes Pascal look bland.

The real story will be what Vega looks like - if its a fast card, it'll cement Pascal as being early, but ultimately a boring stepping to Volta. And if it sucks, its probably the sucking chest wound of AMD's high end GPU business.
 
Pascal's desktop GPUs are more or less Maxwell on a smaller manufacturing process node. And Maxwell is also not that much different from Kepler. Volta will be the first major architectural redesign since Kepler. That's why.
Pascal is still a larger improvement than GCN 1.1-1.4. And there are some architectural improvements, even though they are smaller than the upcoming changes in Volta. Nevertheless, what matters is real world performance, and in that regard Pascal is excellent.
 
Pascal is still a larger improvement than GCN 1.1-1.4. And there are some architectural improvements, even though they are smaller than the upcoming changes in Volta. Nevertheless, what matters is real world performance, and in that regard Pascal is excellent.

The thing is, comparatively we don't know that. As there's no competition in the high end to compare it to. Shrug.
 
Pascal is still a larger improvement than GCN 1.1-1.4. And there are some architectural improvements, even though they are smaller than the upcoming changes in Volta. Nevertheless, what matters is real world performance, and in that regard Pascal is excellent.
That's a highly subjective statement though. Even if you deem Pascal's real world performance excellent, I find it lacking in several metrics.
 
That's a highly subjective statement though. Even if you deem Pascal's real world performance excellent, I find it lacking in several metrics.
Like what precisely? Pascal products is greatly improved over the previous generation (e.g. GTX 1070 vs. GTX 970, …). This is not subjective in any way, it's pure numbers. Pascal is even slightly better at performance per GFlop, which means the architecture is slightly more efficient (regardless of production node).
 
Like what precisely? Pascal products is greatly improved over the previous generation (e.g. GTX 1070 vs. GTX 970, …). This is not subjective in any way, it's pure numbers. Pascal is even slightly better at performance per GFlop, which means the architecture is slightly more efficient (regardless of production node).
Excellent you say? I wouldn't go that far. die shrink gives performance and those Gflops as you said. Smaller size better power consumption and more transistors you can stuff on there. architecturally it doesn't bring much new comparing to Maxwell
 
Excellent you say? I wouldn't go that far. die shrink gives performance and those Gflops as you said. Smaller size better power consumption and more transistors you can stuff on there. architecturally it doesn't bring much new comparing to Maxwell
Please read my post again; I was talking about performance per Gflop (real world performance per theoretical performance). Node shrinks does not have an impact on this.
 
Can still sell 980Ti's second hand here in syd for 550-700+ AUD.... that said I got my gold edition for 400 AUD new so can still make a profit either way :)
 
Ive been Intel/Nvidia Guy for 10 years. Not because im a fan of them just because they built the fastest stuff. Ive never minded paying the extra based on the FPS I get back.

Think this time AMD needs looking at. If they deliver on what they have been promising then this Vega/Rizen might be on par for performance and cost a lot less. I currently have 2x 980Ti's and feel they are worth more than the speculated $200. If they managed to get the 3/4 way SLI working I would happily chuck 2 more in.

If the AMD comes in at 50% the cost of the Intel/Nvidia I'll do a build and see how it goes.
 
Ive been Intel/Nvidia Guy for 10 years. Not because im a fan of them just because they built the fastest stuff. Ive never minded paying the extra based on the FPS I get back.

Think this time AMD needs looking at. If they deliver on what they have been promising then this Vega/Rizen might be on par for performance and cost a lot less. I currently have 2x 980Ti's and feel they are worth more than the speculated $200. If they managed to get the 3/4 way SLI working I would happily chuck 2 more in.

If the AMD comes in at 50% the cost of the Intel/Nvidia I'll do a build and see how it goes.

Chuck away.
3 and 4 way SLI works with 980Ti cards, but is disabled on the next generation. (1070 - 1080 - 1080Ti)
This is assuming you're using a motherboard with enough PCI-E lanes.

I have a pair of 980Ti cards too and they're pretty slick. Now that I finally have a couple of X99 systems, I'm tempted to go with three-way SLI/Crossfire in both of them.
 
Back
Top